{ "features_resolved": { "snapshot_any": true, "snapshot_mem": true, "fork": false, "pause_resume": false, "persistent_disk": true, "whole_disk": false, "persist_scope": true, "volumes": true, "auto_stop_idle": true, "long_sessions": true, "custom_image": true, "image_snapshot": true, "image_oci": true, "vm_isolation": false, "docker_inside": false, "nested_virt": false, "root": null, "systemd": null, "browser": false, "desktop": false, "computer_use": false, "code_interpreter": null, "browser_control": null, "desktop_control": null, "browser_and_desktop": false, "anti_bot": false, "captcha_solving": false, "residential_ip": false, "agent_harnesses": false, "ssh": false, "public_ipv4": false, "inbound_https": true, "custom_domain": true, "raw_tcp_inbound": null, "egress_allowlist": false, "open_internet": null, "static_egress_ip": false, "private_network": true, "self_host": false, "byoc": false, "open_source": false, "soc2": true, "hipaa": true, "sso": false, "gpu": true, "arm64": false, "windows": false, "macos": false, "lang_python": true, "lang_node": false, "lang_go": null, "lang_rust": null, "lang_java": null, "gpu_desktop": null, "credential_injection": null, "egress_http_rules": false, "wake_on_request": true, "live_resize": false, "fork_running": false, "memory_snapshot_fork": false, "webhooks": true, "scheduled_wakeups": null, "audit_logs": true, "eu_data_residency": true, "zero_data_retention": null, "scoped_api_keys": null, "spend_limits": null, "mcp_server": false, "usage_api": true, "snapshot_auto": false, "snapshot_on_demand": true, "harness_api": false, "own_agent_api": false, "fast_boot": null, "strong_isolation": false, "ingress_rules": true, "guest_firewall": false, "secret_proxy": false, "secret_proxy_any": false, "volume_attach": true, "volume_shared": true, "sidecar": null, "fuse": null, "kernel_features": null, "burst": null }, "company": { "id": "cerebrium", "checked": "2026-09-30", "company": "Cerebrium", "ownership": { "value": "independent", "parent": null, "source": "https://www.ycombinator.com/companies/cerebrium", "quote": "Cerebrium is a serverless infrastructure platform for AI applications." }, "stage": { "value": "seed", "source": "https://www.cbinsights.com/company/cerebrium", "quote": "Cerebrium's latest funding round is Seed VC.", "confidence": "medium", "note": "CB Insights lists a Pre-Seed round followed by an $8.5M Seed VC round (led by Gradient Ventures) dated 2025-07-06; treated as seed stage." }, "accelerators": [ { "name": "Y Combinator", "batch": "Winter 2022", "source": "https://www.ycombinator.com/companies/cerebrium", "quote": "batch_name: Winter 2022", "confidence": "high", "via": "research" } ], "total_funding_usd": { "value": 11000000, "status": "estimated", "basis": "CB Insights aggregator total (Pre-Seed + Seed VC rounds); no primary press release found confirming the sum", "source": "https://www.cbinsights.com/company/cerebrium", "quote": "Cerebrium raised a total of $11M." }, "last_round": { "type": "seed", "amount_usd": 8500000, "date": "2025-07", "leads": [ "Gradient Ventures" ], "status": "estimated", "source": "https://www.cbinsights.com/company/cerebrium", "quote": "Last Raised $8.5M | 1 yr ago ... date: 2025-07-06, round: Seed VC, amount: 8.5, investors: Gradient Ventures (lead)" }, "investors": [ "Gradient Ventures", "Authentic Ventures", "Y Combinator", "Maxitech" ], "portfolio_listings": [], "revenue": { "value": null, "status": "unknown" }, "employees": { "value": null, "source": "" }, "note": "YC W22 hint confirmed: same domain (cerebrium.ai) and description as the provider's pricing page. Founded 2021 per CB Insights; HQ listed as Wilmington, DE (likely a registered-agent address, not necessarily where the team operates — founders Michael Louis & Jono are South Africa-based per the HN launch post). No primary-source (press/TechCrunch) confirmation of funding found within the search budget; CB Insights is an aggregator, so funding figures are marked estimated." }, "id": "cerebrium", "name": "Cerebrium", "url": "https://cerebrium.ai/pricing", "category": "gpu-cloud", "isolation": "container", "modes": [ { "key": "interruptible", "label": "Interruptible", "flags": [ "spot", "alt" ], "os": [ "linux" ], "regions": null, "gpu_only": false, "features": {}, "pricing": "resource", "vcpu_h": 0.02358, "ram_gib_h": 0.007992, "cpu_basis": "alloc", "ram_basis": "alloc", "active_floor": 0, "ram_per_vcpu": [ null, null ], "ram_fixed_per_vcpu": null, "vcpu_options": null, "min_vcpu": null, "max_vcpu": 48, "max_ram_gib": 96, "sizes": null, "pool_tiers": null, "requires_always_on": false, "always_on_month_per_instance": null, "commit_note": null, "min_billed_seconds": null, "granularity_s": 1, "start_fee": null, "boot_overhead_s": null, "multiplier": 1, "windows_vcpu_h": null, "gpu": null, "note": "" }, { "key": "protected", "label": "Protected compute", "flags": [ "alt" ], "os": [ "linux" ], "regions": null, "gpu_only": false, "features": {}, "pricing": "resource", "vcpu_h": 0.02358, "ram_gib_h": 0.007992, "cpu_basis": "alloc", "ram_basis": "alloc", "active_floor": 0, "ram_per_vcpu": [ null, null ], "ram_fixed_per_vcpu": null, "vcpu_options": null, "min_vcpu": null, "max_vcpu": 48, "max_ram_gib": 96, "sizes": null, "pool_tiers": null, "requires_always_on": false, "always_on_month_per_instance": null, "commit_note": null, "min_billed_seconds": null, "granularity_s": 1, "start_fee": null, "boot_overhead_s": null, "multiplier": 2, "windows_vcpu_h": null, "gpu": null, "note": "" }, { "key": "serverless-b200-0", "label": "B200, serverless", "flags": [ "spot" ], "os": [ "linux" ], "regions": null, "gpu_only": true, "features": {}, "pricing": "resource", "vcpu_h": 0.02358, "ram_gib_h": 0.007992, "cpu_basis": "alloc", "ram_basis": "alloc", "active_floor": 0, "ram_per_vcpu": [ null, null ], "ram_fixed_per_vcpu": null, "vcpu_options": null, "min_vcpu": null, "sizes": null, "pool_tiers": null, "requires_always_on": false, "always_on_month_per_instance": null, "commit_note": null, "min_billed_seconds": null, "granularity_s": 1, "start_fee": null, "boot_overhead_s": null, "multiplier": 1, "windows_vcpu_h": null, "gpu": { "B200": 6.012 }, "note": "Source: https://www.cerebrium.ai/pricing [grafted from v4-gpu card 2026-09-28]", "product_class": "gpu-instance", "plan_limits_exempt": true, "max_vcpu": 24, "max_ram_gib": 256, "plans": [ "Standard", "Enterprise" ] }, { "key": "serverless-h200-1", "label": "H200, serverless", "flags": [ "spot" ], "os": [ "linux" ], "regions": null, "gpu_only": true, "features": {}, "pricing": "resource", "vcpu_h": 0.02358, "ram_gib_h": 0.007992, "cpu_basis": "alloc", "ram_basis": "alloc", "active_floor": 0, "ram_per_vcpu": [ null, null ], "ram_fixed_per_vcpu": null, "vcpu_options": null, "min_vcpu": null, "sizes": null, "pool_tiers": null, "requires_always_on": false, "always_on_month_per_instance": null, "commit_note": null, "min_billed_seconds": null, "granularity_s": 1, "start_fee": null, "boot_overhead_s": null, "multiplier": 1, "windows_vcpu_h": null, "gpu": { "H200": 4.1976 }, "note": "Source: https://www.cerebrium.ai/pricing [grafted from v4-gpu card 2026-09-28]", "product_class": "gpu-instance", "plan_limits_exempt": true, "max_vcpu": 24, "max_ram_gib": 256, "plans": [ "Standard", "Enterprise" ] }, { "key": "serverless-h100-2", "label": "H100, serverless", "flags": [ "spot" ], "os": [ "linux" ], "regions": null, "gpu_only": true, "features": {}, "pricing": "resource", "vcpu_h": 0.02358, "ram_gib_h": 0.007992, "cpu_basis": "alloc", "ram_basis": "alloc", "active_floor": 0, "ram_per_vcpu": [ null, null ], "ram_fixed_per_vcpu": null, "vcpu_options": null, "min_vcpu": null, "sizes": null, "pool_tiers": null, "requires_always_on": false, "always_on_month_per_instance": null, "commit_note": null, "min_billed_seconds": null, "granularity_s": 1, "start_fee": null, "boot_overhead_s": null, "multiplier": 1, "windows_vcpu_h": null, "gpu": { "H100": 3.3984 }, "note": "Source: https://www.cerebrium.ai/pricing [grafted from v4-gpu card 2026-09-28]", "product_class": "gpu-instance", "plan_limits_exempt": true, "max_vcpu": 24, "max_ram_gib": 256, "plans": [ "Standard", "Enterprise" ] }, { "key": "serverless-rtx-pro-6000-3", "label": "RTX-PRO-6000, serverless", "flags": [ "spot" ], "os": [ "linux" ], "regions": null, "gpu_only": true, "features": {}, "pricing": "resource", "vcpu_h": 0.02358, "ram_gib_h": 0.007992, "cpu_basis": "alloc", "ram_basis": "alloc", "active_floor": 0, "ram_per_vcpu": [ null, null ], "ram_fixed_per_vcpu": null, "vcpu_options": null, "min_vcpu": null, "sizes": null, "pool_tiers": null, "requires_always_on": false, "always_on_month_per_instance": null, "commit_note": null, "min_billed_seconds": null, "granularity_s": 1, "start_fee": null, "boot_overhead_s": null, "multiplier": 1, "windows_vcpu_h": null, "gpu": { "RTX-PRO-6000": 2.4984 }, "note": "Source: https://www.cerebrium.ai/pricing [grafted from v4-gpu card 2026-09-28]. Per-container vCPU/RAM max for RTX-PRO-6000 not listed in docs/hardware/cpu-and-memory.", "product_class": "gpu-instance", "plan_limits_exempt": true, "plans": [ "Standard", "Enterprise" ], "max_vcpu": 24, "max_ram_gib": 218 }, { "key": "serverless-a100-80gb-4", "label": "A100-80GB, serverless", "flags": [ "spot" ], "os": [ "linux" ], "regions": null, "gpu_only": true, "features": {}, "pricing": "resource", "vcpu_h": 0.02358, "ram_gib_h": 0.007992, "cpu_basis": "alloc", "ram_basis": "alloc", "active_floor": 0, "ram_per_vcpu": [ null, null ], "ram_fixed_per_vcpu": null, "vcpu_options": null, "min_vcpu": null, "sizes": null, "pool_tiers": null, "requires_always_on": false, "always_on_month_per_instance": null, "commit_note": null, "min_billed_seconds": null, "granularity_s": 1, "start_fee": null, "boot_overhead_s": null, "multiplier": 1, "windows_vcpu_h": null, "gpu": { "A100-80GB": 2.0988 }, "note": "Source: https://www.cerebrium.ai/pricing [grafted from v4-gpu card 2026-09-28]", "product_class": "gpu-instance", "plan_limits_exempt": true, "max_vcpu": 12, "max_ram_gib": 140, "plans": [ "Standard", "Enterprise" ] }, { "key": "serverless-a100-40gb-5", "label": "A100-40GB, serverless", "flags": [ "spot" ], "os": [ "linux" ], "regions": null, "gpu_only": true, "features": {}, "pricing": "resource", "vcpu_h": 0.02358, "ram_gib_h": 0.007992, "cpu_basis": "alloc", "ram_basis": "alloc", "active_floor": 0, "ram_per_vcpu": [ null, null ], "ram_fixed_per_vcpu": null, "vcpu_options": null, "min_vcpu": null, "sizes": null, "pool_tiers": null, "requires_always_on": false, "always_on_month_per_instance": null, "commit_note": null, "min_billed_seconds": null, "granularity_s": 1, "start_fee": null, "boot_overhead_s": null, "multiplier": 1, "windows_vcpu_h": null, "gpu": { "A100-40GB": 1.998 }, "note": "Source: https://www.cerebrium.ai/pricing [grafted from v4-gpu card 2026-09-28]", "product_class": "gpu-instance", "plan_limits_exempt": true, "max_vcpu": 12, "max_ram_gib": 140, "plans": [ "Standard", "Enterprise" ] }, { "key": "serverless-l40s-6", "label": "L40S, serverless", "flags": [ "spot" ], "os": [ "linux" ], "regions": null, "gpu_only": true, "features": {}, "pricing": "resource", "vcpu_h": 0.02358, "ram_gib_h": 0.007992, "cpu_basis": "alloc", "ram_basis": "alloc", "active_floor": 0, "ram_per_vcpu": [ null, null ], "ram_fixed_per_vcpu": null, "vcpu_options": null, "min_vcpu": null, "sizes": null, "pool_tiers": null, "requires_always_on": false, "always_on_month_per_instance": null, "commit_note": null, "min_billed_seconds": null, "granularity_s": 1, "start_fee": null, "boot_overhead_s": null, "multiplier": 1, "windows_vcpu_h": null, "gpu": { "L40S": 1.9512 }, "note": "Source: https://www.cerebrium.ai/pricing [grafted from v4-gpu card 2026-09-28]. Per-container vCPU/RAM max for L40S not listed in docs/hardware/cpu-and-memory.", "product_class": "gpu-instance", "plan_limits_exempt": true, "max_vcpu": 16, "max_ram_gib": 128 }, { "key": "serverless-a10-7", "label": "A10, serverless", "flags": [ "spot" ], "os": [ "linux" ], "regions": null, "gpu_only": true, "features": {}, "pricing": "resource", "vcpu_h": 0.02358, "ram_gib_h": 0.007992, "cpu_basis": "alloc", "ram_basis": "alloc", "active_floor": 0, "ram_per_vcpu": [ null, null ], "ram_fixed_per_vcpu": null, "vcpu_options": null, "min_vcpu": null, "sizes": null, "pool_tiers": null, "requires_always_on": false, "always_on_month_per_instance": null, "commit_note": null, "min_billed_seconds": null, "granularity_s": 1, "start_fee": null, "boot_overhead_s": null, "multiplier": 1, "windows_vcpu_h": null, "gpu": { "A10": 1.1016 }, "note": "Source: https://www.cerebrium.ai/pricing [grafted from v4-gpu card 2026-09-28]. Per-container vCPU/RAM max for A10 not listed in docs/hardware/cpu-and-memory.", "product_class": "gpu-instance", "plan_limits_exempt": true, "max_vcpu": 48, "max_ram_gib": 192 }, { "key": "serverless-l4-8", "label": "L4, serverless", "flags": [ "spot" ], "os": [ "linux" ], "regions": null, "gpu_only": true, "features": {}, "pricing": "resource", "vcpu_h": 0.02358, "ram_gib_h": 0.007992, "cpu_basis": "alloc", "ram_basis": "alloc", "active_floor": 0, "ram_per_vcpu": [ null, null ], "ram_fixed_per_vcpu": null, "vcpu_options": null, "min_vcpu": null, "sizes": null, "pool_tiers": null, "requires_always_on": false, "always_on_month_per_instance": null, "commit_note": null, "min_billed_seconds": null, "granularity_s": 1, "start_fee": null, "boot_overhead_s": null, "multiplier": 1, "windows_vcpu_h": null, "gpu": { "L4": 0.7992 }, "note": "Source: https://www.cerebrium.ai/pricing [grafted from v4-gpu card 2026-09-28]", "product_class": "gpu-instance", "plan_limits_exempt": true, "max_vcpu": 48, "max_ram_gib": 192 }, { "key": "serverless-t4-9", "label": "T4, serverless", "flags": [ "spot" ], "os": [ "linux" ], "regions": null, "gpu_only": true, "features": {}, "pricing": "resource", "vcpu_h": 0.02358, "ram_gib_h": 0.007992, "cpu_basis": "alloc", "ram_basis": "alloc", "active_floor": 0, "ram_per_vcpu": [ null, null ], "ram_fixed_per_vcpu": null, "vcpu_options": null, "min_vcpu": null, "sizes": null, "pool_tiers": null, "requires_always_on": false, "always_on_month_per_instance": null, "commit_note": null, "min_billed_seconds": null, "granularity_s": 1, "start_fee": null, "boot_overhead_s": null, "multiplier": 1, "windows_vcpu_h": null, "gpu": { "T4": 0.5904 }, "note": "Source: https://www.cerebrium.ai/pricing [grafted from v4-gpu card 2026-09-28]", "product_class": "gpu-instance", "plan_limits_exempt": true, "max_vcpu": 48, "max_ram_gib": 192 }, { "key": "guaranteed-capacity", "label": "Guaranteed burst capacity (min spend)", "flags": [ "sales" ], "pricing": "resource", "commit_note": "Minimum monthly spend commitment, e.g. up to 50 H100s guaranteed for $10,000/month minimum; pay per use above" }, { "key": "serverless-b200-0-protected", "label": "B200, serverless (non-interruptible, 2x)", "flags": [], "os": [ "linux" ], "regions": null, "gpu_only": true, "features": {}, "pricing": "resource", "vcpu_h": 0.04716, "ram_gib_h": 0.015984, "cpu_basis": "alloc", "ram_basis": "alloc", "active_floor": 0, "ram_per_vcpu": [ null, null ], "ram_fixed_per_vcpu": null, "vcpu_options": null, "min_vcpu": null, "sizes": null, "pool_tiers": null, "requires_always_on": false, "always_on_month_per_instance": null, "commit_note": null, "min_billed_seconds": null, "granularity_s": 1, "start_fee": null, "boot_overhead_s": null, "multiplier": 1, "windows_vcpu_h": null, "gpu": { "B200": 12.024 }, "note": "Non-interruptible capacity: Cerebrium docs say it costs 2x the default (interruptible) rate; all rates here are 2x the interruptible mode. Source: https://www.cerebrium.ai/pricing [grafted from v4-gpu card 2026-09-28]", "product_class": "gpu-instance", "plan_limits_exempt": true, "max_vcpu": 24, "max_ram_gib": 256, "plans": [ "Standard", "Enterprise" ] }, { "key": "serverless-h200-1-protected", "label": "H200, serverless (non-interruptible, 2x)", "flags": [], "os": [ "linux" ], "regions": null, "gpu_only": true, "features": {}, "pricing": "resource", "vcpu_h": 0.04716, "ram_gib_h": 0.015984, "cpu_basis": "alloc", "ram_basis": "alloc", "active_floor": 0, "ram_per_vcpu": [ null, null ], "ram_fixed_per_vcpu": null, "vcpu_options": null, "min_vcpu": null, "sizes": null, "pool_tiers": null, "requires_always_on": false, "always_on_month_per_instance": null, "commit_note": null, "min_billed_seconds": null, "granularity_s": 1, "start_fee": null, "boot_overhead_s": null, "multiplier": 1, "windows_vcpu_h": null, "gpu": { "H200": 8.3952 }, "note": "Non-interruptible capacity: Cerebrium docs say it costs 2x the default (interruptible) rate; all rates here are 2x the interruptible mode. Source: https://www.cerebrium.ai/pricing [grafted from v4-gpu card 2026-09-28]", "product_class": "gpu-instance", "plan_limits_exempt": true, "max_vcpu": 24, "max_ram_gib": 256, "plans": [ "Standard", "Enterprise" ] }, { "key": "serverless-h100-2-protected", "label": "H100, serverless (non-interruptible, 2x)", "flags": [], "os": [ "linux" ], "regions": null, "gpu_only": true, "features": {}, "pricing": "resource", "vcpu_h": 0.04716, "ram_gib_h": 0.015984, "cpu_basis": "alloc", "ram_basis": "alloc", "active_floor": 0, "ram_per_vcpu": [ null, null ], "ram_fixed_per_vcpu": null, "vcpu_options": null, "min_vcpu": null, "sizes": null, "pool_tiers": null, "requires_always_on": false, "always_on_month_per_instance": null, "commit_note": null, "min_billed_seconds": null, "granularity_s": 1, "start_fee": null, "boot_overhead_s": null, "multiplier": 1, "windows_vcpu_h": null, "gpu": { "H100": 6.7968 }, "note": "Non-interruptible capacity: Cerebrium docs say it costs 2x the default (interruptible) rate; all rates here are 2x the interruptible mode. Source: https://www.cerebrium.ai/pricing [grafted from v4-gpu card 2026-09-28]", "product_class": "gpu-instance", "plan_limits_exempt": true, "max_vcpu": 24, "max_ram_gib": 256, "plans": [ "Standard", "Enterprise" ] }, { "key": "serverless-rtx-pro-6000-3-protected", "label": "RTX-PRO-6000, serverless (non-interruptible, 2x)", "flags": [], "os": [ "linux" ], "regions": null, "gpu_only": true, "features": {}, "pricing": "resource", "vcpu_h": 0.04716, "ram_gib_h": 0.015984, "cpu_basis": "alloc", "ram_basis": "alloc", "active_floor": 0, "ram_per_vcpu": [ null, null ], "ram_fixed_per_vcpu": null, "vcpu_options": null, "min_vcpu": null, "sizes": null, "pool_tiers": null, "requires_always_on": false, "always_on_month_per_instance": null, "commit_note": null, "min_billed_seconds": null, "granularity_s": 1, "start_fee": null, "boot_overhead_s": null, "multiplier": 1, "windows_vcpu_h": null, "gpu": { "RTX-PRO-6000": 4.9968 }, "note": "Non-interruptible capacity: Cerebrium docs say it costs 2x the default (interruptible) rate; all rates here are 2x the interruptible mode. Source: https://www.cerebrium.ai/pricing [grafted from v4-gpu card 2026-09-28]. Per-container vCPU/RAM max for RTX-PRO-6000 not listed in docs/hardware/cpu-and-memory.", "product_class": "gpu-instance", "plan_limits_exempt": true, "plans": [ "Standard", "Enterprise" ], "max_vcpu": 24, "max_ram_gib": 218 }, { "key": "serverless-a100-80gb-4-protected", "label": "A100-80GB, serverless (non-interruptible, 2x)", "flags": [], "os": [ "linux" ], "regions": null, "gpu_only": true, "features": {}, "pricing": "resource", "vcpu_h": 0.04716, "ram_gib_h": 0.015984, "cpu_basis": "alloc", "ram_basis": "alloc", "active_floor": 0, "ram_per_vcpu": [ null, null ], "ram_fixed_per_vcpu": null, "vcpu_options": null, "min_vcpu": null, "sizes": null, "pool_tiers": null, "requires_always_on": false, "always_on_month_per_instance": null, "commit_note": null, "min_billed_seconds": null, "granularity_s": 1, "start_fee": null, "boot_overhead_s": null, "multiplier": 1, "windows_vcpu_h": null, "gpu": { "A100-80GB": 4.1976 }, "note": "Non-interruptible capacity: Cerebrium docs say it costs 2x the default (interruptible) rate; all rates here are 2x the interruptible mode. Source: https://www.cerebrium.ai/pricing [grafted from v4-gpu card 2026-09-28]", "product_class": "gpu-instance", "plan_limits_exempt": true, "max_vcpu": 12, "max_ram_gib": 140, "plans": [ "Standard", "Enterprise" ] }, { "key": "serverless-a100-40gb-5-protected", "label": "A100-40GB, serverless (non-interruptible, 2x)", "flags": [], "os": [ "linux" ], "regions": null, "gpu_only": true, "features": {}, "pricing": "resource", "vcpu_h": 0.04716, "ram_gib_h": 0.015984, "cpu_basis": "alloc", "ram_basis": "alloc", "active_floor": 0, "ram_per_vcpu": [ null, null ], "ram_fixed_per_vcpu": null, "vcpu_options": null, "min_vcpu": null, "sizes": null, "pool_tiers": null, "requires_always_on": false, "always_on_month_per_instance": null, "commit_note": null, "min_billed_seconds": null, "granularity_s": 1, "start_fee": null, "boot_overhead_s": null, "multiplier": 1, "windows_vcpu_h": null, "gpu": { "A100-40GB": 3.996 }, "note": "Non-interruptible capacity: Cerebrium docs say it costs 2x the default (interruptible) rate; all rates here are 2x the interruptible mode. Source: https://www.cerebrium.ai/pricing [grafted from v4-gpu card 2026-09-28]", "product_class": "gpu-instance", "plan_limits_exempt": true, "max_vcpu": 12, "max_ram_gib": 140, "plans": [ "Standard", "Enterprise" ] }, { "key": "serverless-l40s-6-protected", "label": "L40S, serverless (non-interruptible, 2x)", "flags": [], "os": [ "linux" ], "regions": null, "gpu_only": true, "features": {}, "pricing": "resource", "vcpu_h": 0.04716, "ram_gib_h": 0.015984, "cpu_basis": "alloc", "ram_basis": "alloc", "active_floor": 0, "ram_per_vcpu": [ null, null ], "ram_fixed_per_vcpu": null, "vcpu_options": null, "min_vcpu": null, "sizes": null, "pool_tiers": null, "requires_always_on": false, "always_on_month_per_instance": null, "commit_note": null, "min_billed_seconds": null, "granularity_s": 1, "start_fee": null, "boot_overhead_s": null, "multiplier": 1, "windows_vcpu_h": null, "gpu": { "L40S": 3.9024 }, "note": "Non-interruptible capacity: Cerebrium docs say it costs 2x the default (interruptible) rate; all rates here are 2x the interruptible mode. Source: https://www.cerebrium.ai/pricing [grafted from v4-gpu card 2026-09-28]. Per-container vCPU/RAM max for L40S not listed in docs/hardware/cpu-and-memory.", "product_class": "gpu-instance", "plan_limits_exempt": true, "max_vcpu": 16, "max_ram_gib": 128 }, { "key": "serverless-a10-7-protected", "label": "A10, serverless (non-interruptible, 2x)", "flags": [], "os": [ "linux" ], "regions": null, "gpu_only": true, "features": {}, "pricing": "resource", "vcpu_h": 0.04716, "ram_gib_h": 0.015984, "cpu_basis": "alloc", "ram_basis": "alloc", "active_floor": 0, "ram_per_vcpu": [ null, null ], "ram_fixed_per_vcpu": null, "vcpu_options": null, "min_vcpu": null, "sizes": null, "pool_tiers": null, "requires_always_on": false, "always_on_month_per_instance": null, "commit_note": null, "min_billed_seconds": null, "granularity_s": 1, "start_fee": null, "boot_overhead_s": null, "multiplier": 1, "windows_vcpu_h": null, "gpu": { "A10": 2.2032 }, "note": "Non-interruptible capacity: Cerebrium docs say it costs 2x the default (interruptible) rate; all rates here are 2x the interruptible mode. Source: https://www.cerebrium.ai/pricing [grafted from v4-gpu card 2026-09-28]. Per-container vCPU/RAM max for A10 not listed in docs/hardware/cpu-and-memory.", "product_class": "gpu-instance", "plan_limits_exempt": true, "max_vcpu": 48, "max_ram_gib": 192 }, { "key": "serverless-l4-8-protected", "label": "L4, serverless (non-interruptible, 2x)", "flags": [], "os": [ "linux" ], "regions": null, "gpu_only": true, "features": {}, "pricing": "resource", "vcpu_h": 0.04716, "ram_gib_h": 0.015984, "cpu_basis": "alloc", "ram_basis": "alloc", "active_floor": 0, "ram_per_vcpu": [ null, null ], "ram_fixed_per_vcpu": null, "vcpu_options": null, "min_vcpu": null, "sizes": null, "pool_tiers": null, "requires_always_on": false, "always_on_month_per_instance": null, "commit_note": null, "min_billed_seconds": null, "granularity_s": 1, "start_fee": null, "boot_overhead_s": null, "multiplier": 1, "windows_vcpu_h": null, "gpu": { "L4": 1.5984 }, "note": "Non-interruptible capacity: Cerebrium docs say it costs 2x the default (interruptible) rate; all rates here are 2x the interruptible mode. Source: https://www.cerebrium.ai/pricing [grafted from v4-gpu card 2026-09-28]", "product_class": "gpu-instance", "plan_limits_exempt": true, "max_vcpu": 48, "max_ram_gib": 192 }, { "key": "serverless-t4-9-protected", "label": "T4, serverless (non-interruptible, 2x)", "flags": [], "os": [ "linux" ], "regions": null, "gpu_only": true, "features": {}, "pricing": "resource", "vcpu_h": 0.04716, "ram_gib_h": 0.015984, "cpu_basis": "alloc", "ram_basis": "alloc", "active_floor": 0, "ram_per_vcpu": [ null, null ], "ram_fixed_per_vcpu": null, "vcpu_options": null, "min_vcpu": null, "sizes": null, "pool_tiers": null, "requires_always_on": false, "always_on_month_per_instance": null, "commit_note": null, "min_billed_seconds": null, "granularity_s": 1, "start_fee": null, "boot_overhead_s": null, "multiplier": 1, "windows_vcpu_h": null, "gpu": { "T4": 1.1808 }, "note": "Non-interruptible capacity: Cerebrium docs say it costs 2x the default (interruptible) rate; all rates here are 2x the interruptible mode. Source: https://www.cerebrium.ai/pricing [grafted from v4-gpu card 2026-09-28]", "product_class": "gpu-instance", "plan_limits_exempt": true, "max_vcpu": 48, "max_ram_gib": 192 } ], "storage": { "included_disk_gib": null, "disk_gib_month": 0.05, "disk_billed_when_stopped": true, "snapshot_gib_month": null, "paused_ram_billed": null }, "network": { "egress_free_gib": null, "egress_gib": null, "ipv4_month": null, "ipv4_available": null }, "plans": [ { "name": "Hobby", "fee": 0, "flags": [], "fee_is_credit": false, "included_usd": 0, "per_seat": null, "concurrency": 500, "max_session_h": null, "max_vcpu": null, "max_ram_gib": null, "sizes_allowed": null, "note": "3 seats, up to 3 deployed apps, 5 concurrent GPUs, only T4/A10/L4/L40S/INF2/CPU. Log retention 7 days (table) vs 1 day (plan card). No audit log." }, { "name": "Standard", "fee": 100, "flags": [], "fee_is_credit": false, "included_usd": 0, "per_seat": null, "concurrency": 1000, "max_session_h": null, "max_vcpu": null, "max_ram_gib": null, "sizes_allowed": null, "note": "Seats: 10 (comparison table); plan card says 'Unlimited seats'. No per-seat fee. Unlimited deployed apps. Required for A100/H100/H200/B200/RTX PRO 6000. Audit log 7 days. Log retention 30 days." }, { "name": "Enterprise", "fee": null, "flags": [ "sales" ], "fee_is_credit": false, "included_usd": null, "per_seat": null, "concurrency": null, "max_session_h": null, "max_vcpu": null, "max_ram_gib": null, "sizes_allowed": null, "note": "Volume discounts; discounts for longer-term commitments (sales-quoted); unlimited GPU concurrency; TRN1 (Trainium) only on Enterprise." } ], "free": { "monthly_credit": null, "one_time_credit": null }, "features": { "linux": true, "windows": false, "macos": false, "arm64": false, "gpu": true, "snapshot": "mem", "pause_resume": false, "fork": false, "persistent_disk": true, "volumes": true, "auto_stop_idle": true, "max_session_h": null, "custom_image": true, "docker_inside": false, "nested_virt": false, "root": null, "systemd": null, "browser": false, "desktop": false, "computer_use": false, "code_interpreter": null, "agent_harnesses": false, "ssh": false, "public_ipv4": false, "inbound_https": true, "custom_domain": true, "raw_tcp_inbound": null, "egress_allowlist": false, "static_egress_ip": false, "private_network": true, "self_host": false, "byoc": false, "open_source": false, "soc2": true, "hipaa": true, "sso": false, "gpu_types": [ "B200", "H200", "H100", "RTX PRO 6000", "A100-80GB", "A100-40GB", "L40S", "A10", "L4", "T4", "RTX-PRO-6000" ], "regions": [ "us", "eu", "asia", "other" ], "lang_python": true, "lang_node": false, "lang_go": null, "lang_rust": null, "lang_java": null, "gpu_desktop": null, "credential_injection": null, "egress_http_rules": false, "wake_on_request": true, "live_resize": false, "fork_running": false, "memory_snapshot_fork": false, "webhooks": true, "scheduled_wakeups": null, "audit_logs": true, "eu_data_residency": true, "zero_data_retention": null, "scoped_api_keys": null, "spend_limits": null, "mcp_server": false, "snapshot_auto": false, "snapshot_on_demand": true, "harness_api": false, "ingress_rules": true, "secret_proxy": false, "volume_attach": true, "volume_shared": true, "isolation": "container", "image_oci": true, "anti_bot": false, "captcha_solving": false, "egress_ip_type": "datacenter", "own_agent_api": false, "guest_firewall": false, "max_session_unlimited": true, "usage_api": true, "persist_scope": "folder", "persist_path": "/persistent-storage" }, "perf": { "cpu_runs_s": null, "cold_start_ms": null, "burst_tti_ms": null, "dax_s": null }, "caveats": [ "CPU $0.00000655/vCPU-s, RAM $0.00000222/GB-s. Hardware docs say actual CPU/memory usage; pricing FAQ also says resources allocated while processing. Record utilization ambiguity rather than promise savings.", "Default interruptible tier can be preempted or consolidated. Protected tier costs 2× all compute including GPU, CPU and RAM; storage unaffected.", "Storage first 100 GB free then $0.05/GB-month. Standard $100 is plus compute (no stated credits). Pricing comparison conflicts on Standard seats and Hobby log retention.", "GPU listed per-second rates converted without rounding to marketing hourly approximations.", "Null values mean unknown, not free/unlimited. No model-token cost, tax or unverified ancillary charge included.", "Native browser/device/task/credit prices with unknown CPU/RAM are deliberately non-priceable. Check regime before interpreting any partial resource estimate.", "GPU modes grafted from the v4-gpu research card (per-GPU rates; exact rows in research/gpu.json).", "GPU concurrency per plan (pricing page 2026-09-28): Hobby 5 GPUs, Standard ($100/mo) 30 GPUs, Enterprise unlimited; plan.concurrency holds the CPU container limits (500 / 1,000). GPU cap not enforced by the engine.", "Per-container max (docs/hardware/cpu-and-memory): H100/H200/B200 24 vCPU / 256 GB; A100 12 vCPU / 140 GB; L4/T4 48 vCPU / 192 GB; L40S/A10/RTX PRO 6000 not listed.", "AWS Inferentia2 (INF2, Hobby+) and Trainium (TRN1, Enterprise) are offered but have no public price; TRN1 containers up to 128 vCPU / 512 GB. Max 8 GPUs per container for every GPU type.", "Builds are billed at the app's compute rate when requirements/code change (cached steps cheaper). min_replicas>0 bills 24/7.", "Memory+GPU checkpointing (snapshots for cold start) is early beta; no separate storage price published.", "Platform cold-start time is not billed; user initialization code outside the request function (e.g. loading a model into GPU memory) is billed.", "Persistent storage: one volume per region per project (mounted /persistent-storage), resizable 50 GB-1 TB self-serve; global apps also get a fixed 1 TB /global-persistent-storage volume. Pricing page: first 100 GB free then $0.05/GB-mo; whether billing is on provisioned size or used bytes is not stated.", "No container lifetime cap, but each request times out at response_grace_period (default 15 min, raise in cerebrium.toml) and async runs max out at 12 h. Interruptible and protected instances alike can be drained (SIGTERM, then SIGKILL after the grace period) and moved to a new node for maintenance; protected only avoids preemption and consolidation.", "Cerebrium does not say which isolation layer (runc, gVisor or microVM) runs customer containers; 'container' is inferred from the Dockerfile deployment model." ], "sources": [ "https://www.ycombinator.com/companies/cerebrium", "https://cerebrium.ai/pricing", "https://cerebrium.ai/docs/hardware/cpu-and-memory", "https://cerebrium.ai/docs/llms-full.txt", "https://cerebrium.ai/", "https://cerebrium.ai/docs/getting-started/introduction", "https://cerebrium.ai/llms.txt", "https://www.cerebrium.ai/pricing", "https://cerebrium.ai/docs/hardware/cpu-and-memory.md", "https://docs.cerebrium.ai/llms.txt", "https://cerebrium.ai/docs/endpoints/async", "https://cerebrium.ai/docs/scaling/scaling-apps", "https://cerebrium.ai/docs/toml-reference/toml-reference", "https://cerebrium.ai/resources/sandboxed-agent-code-execution-serverless-gpu" ], "fleet": { "default_max_instances": 500, "default_max_vcpu": null, "default_max_gpus": 5, "raise": "self-serve", "raise_time_days": null, "verification": null, "tiers": "Hobby: 500 CPU containers, 5 concurrent GPUs, 3 deployed apps. Standard: 1,000 CPU containers, 30 GPU concurrency. Enterprise custom/unlimited advertised.", "notes": "Instance field is CPU-container concurrency, not 500 GPU machines. Public plan ceilings do not reserve stock; new-account payment/KYC checks and approval time not specified.", "sources": [ "https://cerebrium.ai/pricing" ], "researched": "2026-09-28" }, "audit_2026_09_29": { "applied": 21, "source": "checked against the provider's live pricing and docs pages", "reviewed": 17 }, "feature_evidence": { "snapshot_auto": { "value": false, "source": "https://cerebrium.ai/docs/performance/checkpointing.md", "quote": "Ephemeral filesystem: Any files written to disk before the checkpoint are not copied to the restored container. Only memory is checkpointed.", "note": "Serverless containers are stateless: scale-down after `cooldown` sends SIGTERM then SIGKILL (scaling-apps) with no state saved. The only checkpoint is the explicit, app-triggered one below; nothing is saved automatically on stop/idle.", "checked": "2026-09-29" }, "snapshot_on_demand": { "value": true, "source": "https://cerebrium.ai/docs/performance/checkpointing.md", "quote": "Memory checkpointing takes a snapshot of a container's CPU memory and GPU memory, and uses it to speed up the startup of future containers.", "note": "Early beta, opt-in via `[cerebrium.experimental] checkpointing = true`. Triggered by the app itself with a POST to the in-container sidecar http://169.254.169.253:8234/checkpoint (no external API/CLI). CPU+GPU memory included, filesystem NOT included; one checkpoint per build version ('skips checkpoint creation if: A checkpoint already exists for the current build version'); all later containers of that deployment restore from it (a cold-start template, not a user-managed fork/restore of a specific container).", "checked": "2026-09-29" }, "harness_api": { "value": false, "source": "https://cerebrium.ai/llms.txt", "quote": "Cerebrium is serverless GPU infrastructure for real-time AI — voice agents, video models, LLMs, and custom ML apps. Sub-second cold starts, pay-per-second billing, no Kubernetes.", "note": "Cerebrium is a serverless GPU inference platform for deploying your own Python/container workloads (voice, video, LLM, ML), not a coding-agent sandbox product. No coding-harness (Claude Code/Codex/OpenCode) API exists anywhere in docs, llms.txt, or the site; the only agent-adjacent content is a blog post on sandboxing agent-generated code, not a first-party harness-driving API. Confirms false.", "checked": "2026-09-30" }, "ingress_rules": { "value": true, "source": "https://cerebrium.ai/docs/toml-reference/toml-reference.md", "quote": "`disable_auth` defaults to `true`, so omitting it leaves app endpoints callable without a token. Set `disable_auth = false` to require the JWT token", "note": "Endpoints are public by default; per-app toggle to require a JWT bearer token. Functions prefixed with '_' are not exposed ('unless marked private by prefixing the function name with an underscore', inference-api). Custom runtime exposes one configured `port`. Inter-cluster routing keeps app-to-app traffic in-region, still authenticated unless auth disabled. No IP allowlists documented.", "checked": "2026-09-29" }, "secret_proxy": { "value": false, "source": "https://cerebrium.ai/docs/other-topics/using-secrets.md", "quote": "Each secret is exposed to the app as an environment variable.", "note": "Project- or app-level secrets, AES-256 at rest, loaded as env vars at container start (restart needed after change). No credential-injecting proxy.", "checked": "2026-09-29" }, "volume_attach": { "value": true, "source": "https://cerebrium.ai/docs/storage/managing-files.md", "quote": "Cerebrium offers file management through a persistent volume that's available to all apps in a project. This storage mounts at `/persistent-storage`", "note": "Not arbitrary attachable volumes: one default volume per region per project (id `default-`), resizable 50 GB-1 TB via dashboard/API (no shrink); global apps also get a fixed 1 TB /global-persistent-storage. Pricing: first 100 GB free then $0.05/GB-month (card).", "checked": "2026-09-29" }, "volume_shared": { "value": true, "source": "https://cerebrium.ai/docs/storage/managing-files.md", "quote": "It is a single volume shared across every region the app runs in: files written there are visible from all regions, and reads are cached per region.", "note": "Read-write. The regional /persistent-storage is mounted by all apps/containers of the project at once; the global volume is shared across regions. Consistency/locking semantics not documented.", "checked": "2026-09-29" }, "isolation": { "value": "container", "source": "https://cerebrium.ai/docs/container-images/custom-dockerfiles", "quote": "Cerebrium supports deploying existing containerized apps — from standard Python apps to compiled Rust binaries — using a custom Dockerfile.", "note": "Serverless container platform built from Dockerfiles; 'container' is a reading of that model. Cerebrium does not state which isolation layer it runs customer containers under (runc, gVisor or microVM) in its docs or its sandboxing resource page, so per-customer-VM vs shared-kernel is unconfirmed.", "checked": "2026-10-14" }, "image_oci": { "value": true, "source": "https://cerebrium.ai/docs/container-images/custom-dockerfiles", "quote": "Cerebrium supports deploying existing containerized apps ... using a custom Dockerfile. This allows portable, locally reproducible deployment environments.", "note": "Custom Dockerfile / custom base image (docker_base_image_url) and private registry (Docker Hub, ECR) support.", "checked": "2026-09-30" }, "docker_inside": { "value": false, "source": "https://cerebrium.ai/docs/container-images/custom-dockerfiles", "quote": "Cerebrium uses a single cerebrium.toml file. The system handles container lifecycle, networking, and scaling automatically.", "note": "not offered: outside the product's scope — apps run as a single managed container (built from your Dockerfile); no documented ability to run Docker-in-Docker or start further containers inside the running app.", "checked": "2026-09-30" }, "nested_virt": { "value": false, "source": "https://cerebrium.ai/docs/container-images/custom-dockerfiles", "quote": "Cerebrium supports deploying existing containerized apps — from standard Python apps to compiled Rust binaries.", "note": "not offered: outside the product's scope — a serverless container platform for Python/custom-runtime inference apps (underlying isolation layer not stated by Cerebrium); no /dev/kvm or nested-VM capability is documented.", "checked": "2026-10-14" }, "desktop": { "value": false, "source": "https://cerebrium.ai/pricing", "quote": "Deploy AI infrastructure that scales automatically ... serverless GPU compute for inference and training.", "note": "not offered: outside the product's scope — a serverless inference/API platform, not a desktop/GUI product.", "checked": "2026-09-30" }, "browser": { "value": false, "source": "https://cerebrium.ai/pricing", "quote": "Deploy AI infrastructure that scales automatically ... serverless GPU compute for inference and training.", "note": "not offered: outside the product's scope — no managed/real browser product; docs only cover an example blog post on sandboxing agent code, not a browser feature.", "checked": "2026-09-30" }, "computer_use": { "value": false, "source": "https://cerebrium.ai/pricing", "quote": "Deploy AI infrastructure that scales automatically ... serverless GPU compute for inference and training.", "note": "not offered: outside the product's scope — no screenshot/click desktop-driving API; platform exposes REST/WebSocket endpoints for user-authored inference functions only.", "checked": "2026-09-30" }, "anti_bot": { "value": false, "source": "https://cerebrium.ai/pricing", "quote": "Deploy AI infrastructure that scales automatically ... serverless GPU compute for inference and training.", "note": "not offered: outside the product's scope — an ML inference/compute platform, not a browser-automation/scraping product.", "checked": "2026-09-30" }, "captcha_solving": { "value": false, "source": "https://cerebrium.ai/pricing", "quote": "Deploy AI infrastructure that scales automatically ... serverless GPU compute for inference and training.", "note": "not offered: outside the product's scope — no CAPTCHA-solving feature documented; not a browser/scraping product.", "checked": "2026-09-30" }, "egress_ip_type": { "value": "datacenter", "source": "https://cerebrium.ai/pricing", "quote": "Deploy AI infrastructure that scales automatically ... serverless GPU compute for inference and training.", "note": "Containers run on Cerebrium's own cloud/GPU node fleet; outbound traffic is standard cloud-datacenter IPs. No residential/mobile proxy or static egress IP product documented.", "checked": "2026-09-30" }, "ssh": { "value": false, "source": "https://cerebrium.ai/docs/container-images/custom-dockerfiles", "quote": "Cerebrium uses a single cerebrium.toml file. The system handles container lifecycle, networking, and scaling automatically based on this configuration.", "note": "not offered: outside the product's scope — a fully managed serverless container platform; no SSH access into running containers is documented anywhere in the docs/CLI reference.", "checked": "2026-09-30" }, "snapshot": { "value": "mem", "source": "https://cerebrium.ai/docs/performance/checkpointing", "quote": "Memory checkpointing takes a snapshot of a container's CPU memory and GPU memory, and uses it to speed up the startup of future containers.", "note": "Early-beta, opt-in, app-triggered checkpoint of CPU+GPU memory (not filesystem) used to seed future cold starts of the same build — not a general user-directed snapshot/restore of a specific running container.", "checked": "2026-09-30" }, "fork": { "value": false, "source": "https://cerebrium.ai/docs/performance/checkpointing", "quote": "Only memory is checkpointed ... uses it to speed up the startup of future containers.", "note": "not offered: no documented ability to clone a running machine into a new independent machine on demand; the checkpoint feature seeds future cold starts of the same deployment, not user-directed forking.", "checked": "2026-09-30" }, "fork_running": { "value": false, "source": "https://cerebrium.ai/docs/performance/checkpointing", "quote": "Memory checkpointing takes a snapshot of a container's CPU memory and GPU memory, and uses it to speed up the startup of future containers.", "note": "not offered: no API/CLI to fork a currently running container without stopping it.", "checked": "2026-09-30" }, "memory_snapshot_fork": { "value": false, "source": "https://cerebrium.ai/docs/performance/checkpointing", "quote": "Memory checkpointing takes a snapshot of a container's CPU memory and GPU memory, and uses it to speed up the startup of future containers.", "note": "not offered as a fork primitive: the memory checkpoint speeds up cold starts of new containers of the same build; it is not exposed as 'clone this running machine, memory and all, into a new independent one'.", "checked": "2026-09-30" }, "pause_resume": { "value": false, "source": "https://cerebrium.ai/pricing", "quote": "Deploy AI infrastructure that scales automatically ... serverless GPU compute for inference and training.", "note": "not offered: outside the product's scope — serverless containers scale to zero and cold-start again; there is no user-facing pause/resume of a specific running instance, only autoscaling (min_replicas/cooldown).", "checked": "2026-09-30" }, "wake_on_request": { "value": true, "source": "https://cerebrium.ai/docs/llms-full.txt", "quote": "A request enters through the region closest to where it is made, then routes to a region with a warm instance before a cold start is triggered elsewhere.", "note": "Core serverless model: scaled-to-zero apps cold-start automatically on an incoming request.", "checked": "2026-09-30" }, "persistent_disk": { "value": true, "source": "https://cerebrium.ai/docs/storage/managing-files", "quote": "Cerebrium offers file management through a persistent volume that's available to all apps in a project. This storage mounts at `/persistent-storage`", "note": "Persistent volume(s) per project/region survive container restarts/redeploys; container root filesystem itself is ephemeral per the checkpointing docs.", "checked": "2026-09-30" }, "live_resize": { "value": false, "source": "https://cerebrium.ai/docs/hardware/cpu-and-memory", "quote": "CPU resources are specified as vCPU units (float) in the `cerebrium.toml` file", "note": "not offered: vCPU/RAM/GPU are set in the app's cerebrium.toml config and applied via redeploy/rebuild; no documented API to resize CPU/RAM of an already-running container in place.", "checked": "2026-09-30" }, "agent_harnesses": { "value": false, "source": "https://cerebrium.ai/pricing", "quote": "Deploy AI infrastructure that scales automatically ... serverless GPU compute for inference and training.", "note": "not offered: outside the product's scope — no preinstalled/templated coding-agent harnesses (Claude Code, Codex, OpenCode); templates in docs are for ML inference/voice-agent apps.", "checked": "2026-09-30" }, "own_agent_api": { "value": false, "source": "https://cerebrium.ai/pricing", "quote": "Deploy AI infrastructure that scales automatically ... serverless GPU compute for inference and training.", "note": "not offered: outside the product's scope — Cerebrium runs customer-authored inference/agent code; it does not ship its own first-party browser/computer-use/research agent product.", "checked": "2026-09-30" }, "mcp_server": { "value": false, "source": "https://cerebrium.ai/pricing", "quote": "Deploy AI infrastructure that scales automatically ... serverless GPU compute for inference and training.", "note": "not offered: no MCP server for controlling Cerebrium sandboxes found in docs/llms-full.txt or GitHub org.", "checked": "2026-09-30" }, "lang_python": { "value": true, "source": "https://cerebrium.ai/docs/llms-full.txt", "quote": "The Python runtime version forms the foundation of every Cerebrium app. Supported versions: 3.10 to 3.13.", "note": "Python is the primary first-party runtime with its own dependency management in cerebrium.toml.", "checked": "2026-09-30" }, "lang_node": { "value": false, "source": "https://cerebrium.ai/docs/container-images/custom-dockerfiles", "quote": "Cerebrium supports deploying existing containerized apps — from standard Python apps to compiled Rust binaries — using a custom Dockerfile.", "note": "No first-party Node.js runtime/SDK; Node (or any other language) apps are only reachable indirectly via a custom Dockerfile/custom runtime, same as Rust — not offered out of the box.", "checked": "2026-09-30" }, "public_ipv4": { "value": false, "source": "https://cerebrium.ai/docs/toml-reference/toml-reference", "quote": "`disable_auth` defaults to `true`, so omitting it leaves app endpoints callable without a token.", "note": "Apps are reached via Cerebrium-issued HTTPS endpoints/custom domains (inbound_https=true in card), not a dedicated public IPv4 per machine; no per-container IPv4 documented.", "checked": "2026-09-30" }, "egress_allowlist": { "value": false, "source": "https://cerebrium.ai/pricing", "quote": "Deploy AI infrastructure that scales automatically ... serverless GPU compute for inference and training.", "note": "No provider-side outbound allow/deny list or no-internet mode documented anywhere in docs/llms-full.txt or toml-reference.", "checked": "2026-09-30" }, "egress_http_rules": { "value": false, "source": "https://cerebrium.ai/pricing", "quote": "Deploy AI infrastructure that scales automatically ... serverless GPU compute for inference and training.", "note": "not offered: no HTTP method/path-level egress rules documented; egress_allowlist itself is not offered.", "checked": "2026-09-30" }, "guest_firewall": { "value": false, "source": "https://cerebrium.ai/docs/container-images/custom-dockerfiles", "quote": "Cerebrium supports deploying existing containerized apps — from standard Python apps to compiled Rust binaries.", "note": "Deployments run as managed serverless containers (underlying isolation layer not stated by Cerebrium); no iptables/nftables-in-guest capability documented, and root/sudo access is not exposed.", "checked": "2026-10-14" }, "static_egress_ip": { "value": false, "source": "https://cerebrium.ai/pricing", "quote": "Deploy AI infrastructure that scales automatically ... serverless GPU compute for inference and training.", "note": "No fixed/static outbound IP product documented.", "checked": "2026-09-30" }, "windows": { "value": false, "source": "https://cerebrium.ai/pricing", "quote": "Deploy AI infrastructure that scales automatically ... serverless GPU compute for inference and training.", "note": "not offered: outside the product's scope — Linux-only container runtime (Windows is only an OS for the local CLI tool, not a machine you deploy to).", "checked": "2026-09-30" }, "macos": { "value": false, "source": "https://cerebrium.ai/pricing", "quote": "Deploy AI infrastructure that scales automatically ... serverless GPU compute for inference and training.", "note": "not offered: outside the product's scope — Linux-only container runtime (macOS is only supported for installing the local CLI via Homebrew, not as a machine OS).", "checked": "2026-09-30" }, "arm64": { "value": false, "source": "https://cerebrium.ai/docs/container-images/private-docker-registry", "quote": "Your Docker images MUST support `linux/amd64` architecture. This is required", "note": "Platform requires linux/amd64 images; no ARM64 machine option offered.", "checked": "2026-09-30" }, "self_host": { "value": false, "source": "https://cerebrium.ai/pricing", "quote": "Deploy AI infrastructure that scales automatically ... serverless GPU compute for inference and training.", "note": "not offered: outside the product's scope — Cerebrium is a hosted managed cloud platform; no self-hosted/on-prem deployment of the Cerebrium engine is documented for customers.", "checked": "2026-09-30" }, "byoc": { "value": false, "source": "https://cerebrium.ai/pricing", "quote": "Deploy AI infrastructure that scales automatically ... serverless GPU compute for inference and training.", "note": "not offered: no documented option to deploy the managed Cerebrium service into a customer's own cloud account.", "checked": "2026-09-30" }, "open_source": { "value": false, "source": "https://github.com/CerebriumAI/cerebrium", "quote": "CerebriumAI/cerebrium — releases include the Cerebrium CLI binaries (cerebrium_linux_amd64.deb, cerebrium_cli_windows_amd64.zip).", "note": "The public CerebriumAI GitHub org ships the CLI and example apps, but the core serverless engine/platform itself is closed-source and proprietary.", "checked": "2026-09-30" }, "max_session_h": { "value": "none", "source": "https://cerebrium.ai/docs/scaling/scaling-apps", "quote": "During normal replica operation, this acts as a request timeout value.", "note": "No documented cap on how long a container/replica lives (min_replicas keeps instances warm 24/7), so the container-level value stays 'none'. Each request is bounded by `response_grace_period` (toml default 900 s = 15 min, no maximum stated); async runs are capped: 'Async functions run for a maximum of **12 hours**, bounded by the `response_grace_period` in `cerebrium.toml`. This defaults to 15 minutes.' (docs/endpoints/async). No instance is pinned to one node: 'Both interruptible and protected instances can be migrated to a new node when the underlying node is cycled for maintenance. In every case the instance receives SIGTERM and is drained within `response_grace_period` before SIGKILL.' (scaling-apps).", "checked": "2026-10-14" }, "sso": { "value": false, "source": "https://www.cerebrium.ai/pricing", "quote": "Detailed Plan Comparison Hobby Standard Enterprise ... Seats 3 10 Custom ... SOC2 compliance ... HIPPA, GDPR, ISO 27001 ... Dedicated Slack support White glove onboarding ML engineering services", "note": "The pricing page's detailed Hobby/Standard/Enterprise comparison table lists seats, log retention, SOC2/HIPAA/GDPR/ISO27001, domains, concurrency and support tiers but has no SSO/SAML/OIDC line item anywhere, even for Enterprise; the security & data privacy docs page (SOC2, HIPAA, GDPR, ISO, MFA, Vanta) and the trust center also do not mention SSO. No first-party page advertises SAML/OIDC single sign-on for the Cerebrium dashboard.", "checked": "2026-09-30" }, "usage_api": { "value": true, "source": "https://cerebrium.ai/docs/api-reference/apps/get-an-apps-request-and-cost-breakdown-over-time", "quote": "How much traffic an app served and what it cost to run, over a chosen period. Requests are counted per time bucket and split by response status (2xx/4xx/5xx, normal and error WebSocket closes, and anything else); cost is split across GPU, CPU and memory.", "note": "granularity: per-app; cost: yes; GET /v2/projects/{project_id}/apps/{app_id}/dashboard-metrics endpoint returns billable usage (cost breakdown by GPU/CPU/memory) over time per app. Also GET /v2/projects/{project_id}/apps/{app_id}/cost for cost summary per app, and GET /v2/projects/{project_id}/cost for project-level cost.", "confidence": "high", "checked": "2026-10-01" }, "persist_scope": { "value": "folder", "source": "https://cerebrium.ai/docs/storage/managing-files", "quote": "Cerebrium offers file management through a persistent volume that's available to all apps in a project. This storage mounts at `/persistent-storage`", "note": "Only /persistent-storage survives a stop; the rest of the filesystem (packages installed elsewhere) resets.", "checked": "2026-10-01" } }, "locations": [ { "country": "US", "state": "VA", "city": "Ashburn", "provider_region": "us-east-1", "self_serve": true, "choosable": true }, { "country": "US", "state": "MO", "city": "Kansas City", "provider_region": "us-central1", "self_serve": true, "choosable": true }, { "country": "SE", "state": null, "city": "Stockholm", "provider_region": "eu-north-1", "self_serve": true, "choosable": true }, { "country": "FI", "state": null, "city": "Helsinki", "provider_region": "eu-north1", "self_serve": true, "choosable": true }, { "country": "US", "state": "OR", "city": "Portland", "provider_region": "us-west-2", "self_serve": false, "choosable": true }, { "country": "GB", "state": null, "city": "London", "provider_region": "eu-west-2", "self_serve": false, "choosable": true }, { "country": "DE", "state": null, "city": "Frankfurt", "provider_region": "eu-central-1", "self_serve": false, "choosable": true }, { "country": "IN", "state": null, "city": "Mumbai", "provider_region": "ap-south-1", "self_serve": false, "choosable": true }, { "country": "JP", "state": null, "city": "Tokyo", "provider_region": "ap-northeast-1", "self_serve": false, "choosable": true }, { "country": "BR", "state": null, "city": "São Paulo", "provider_region": "sa-east-1", "self_serve": false, "choosable": true }, { "country": "CA", "state": "QC", "city": "Montreal", "provider_region": "ca-central-1", "self_serve": false, "choosable": true }, { "country": "AE", "state": null, "city": "Dubai", "provider_region": "me-central-1", "self_serve": false, "choosable": true } ], "locations_source": "https://cerebrium.ai/docs/deployments/multi-region-deployment", "locations_quote": "us-east-1 (N. Virginia), us-central1 (Kansas City), eu-north-1 (Stockholm), eu-north1 (Finland) are generally available. Contact support to enable access to us-west-2 (Oregon), eu-west-2 (United Kingdom), eu-central-1 (Frankfurt), ap-south-1 (Mumbai), ap-northeast-1 (Tokyo), sa-east-1 (São Paulo), c", "locations_checked": "2026-10-01" }